On Sat, Nov 22, 2008 at 03:05:22PM -0700, Scott Long wrote: > A neat hack would be for the kernel linker to scan the text and do a > drop-in replacement of the opcode that is appropriate for the platform. > I can't see how a CPU_XXX definition would work because it's just a > compile time construct, one that can be included with any kernel > compile.
Yes, it is possible to do that. Less drastic change is to directly
check features. I moved slow code to separate section to eliminate
unconditional jump in fast path.
Only compile-tested.
diff --git a/sys/conf/ldscript.i386 b/sys/conf/ldscript.i386
index a94f32f..49d9636 100644
--- a/sys/conf/ldscript.i386
+++ b/sys/conf/ldscript.i386
@@ -45,6 +45,7 @@ SECTIONS
.text :
{
*(.text)
+ *(.text.offpath)
*(.stub)
/* .gnu.warning sections are handled specially by elf32.em. */
*(.gnu.warning)
diff --git a/sys/i386/include/atomic.h b/sys/i386/include/atomic.h
index f6bcf0c..5806d41 100644
--- a/sys/i386/include/atomic.h
+++ b/sys/i386/include/atomic.h
@@ -32,11 +32,38 @@
#error this file needs sys/cdefs.h as a prerequisite
#endif
-
+#ifdef _KERNEL
#if defined(I686_CPU)
-#define mb() __asm__ __volatile__ ("mfence;": : :"memory")
-#define wmb() __asm__ __volatile__ ("sfence;": : :"memory")
-#define rmb() __asm__ __volatile__ ("lfence;": : :"memory")
+#define mb() __asm__ __volatile__ ( \
+ "testl\t$0x04000000,cpu_feature\n" \
+ "\tje\t2f\n" \
+ "\tmfence\n" \
+ "1:\n" \
+ "\t.section\t.text.offpath\n" \
+ "2:\tlock;addl\t$0,cpu_feature\n" \
+ "\tjmp\t1b\n" \
+ "\t.text" \
+ : : :"memory")
+#define wmb() __asm__ __volatile__ ( \
+ "testl\t$0x02000000,cpu_feature\n" \
+ "\tje\t2f\n" \
+ "\tsfence\n" \
+ "1:\n" \
+ "\t.section\t.text.offpath\n" \
+ "2:\tlock;addl\t$0,cpu_feature\n" \
+ "\tjmp\t1b\n" \
+ "\t.text" \
+ : : :"memory")
+#define rmb() __asm__ __volatile__ ( \
+ "testl\t$0x04000000,cpu_feature\n" \
+ "\tje\t2f\n" \
+ "\tlfence\n" \
+ "1:\n" \
+ "\t.section\t.text.offpath\n" \
+ "2:\tlock;addl\t$0,cpu_feature\n" \
+ "\tjmp\t1b\n" \
+ "\t.text" \
+ : : :"memory")
#else
/*
* do we need a serializing instruction?
@@ -45,7 +72,7 @@
#define wmb()
#define rmb()
#endif
-
+#endif
/*
* Various simple operations on memory, each of which is atomic in the
pgpnsZga0lo9O.pgp
Description: PGP signature
